Figure 2.

Bayesian model averaging strategies On the left is the model averaging protocol of [4], on the right is our proposed model averaging strategy. The three key differences are the choice of scoring function, learning multiple models from each bootstrap resample, and automatic determination of the probability threshold.

Broom et al. BMC Bioinformatics 2012 13(Suppl 13):S10   doi:10.1186/1471-2105-13-S13-S10